Importing neccessary modules

Reading and modefying data

Correlating Data

Normal

Pearson when 9300 observations are correlating

Kendal when 9300 observations are correlating

Spearman when 9300 observations are correlating

By looking above result spearman is better than other 3 methods

Now as seen above All the features have less correlation with target values

Data (max,min,avg)

Dropping AggregatedTrustValue from data and modifying outranged Rows sum to 1

Classifying data as Trustworthy or Untrustworthy

Classifying Trustworthy and Untrustworthy rows

Datatypes of varibles in modified data

Features we need for modelling

Splitting data and Training and Testing

Decision Tree

RandomForest Classifier

Bagging

Adaboost Classifer

Voting Classifier

Model Comparisons

End

SUM Represents AggregatedTrustValue